423A35C7/pytorch3d
1
0
Fork 0
mirror of https://github.com/facebookresearch/pytorch3d.git synced 2025-11-04 18:02:14 +08:00
Code Issues Packages Projects Releases Wiki Activity

more code blocks for readthedocs

Browse Source 
Summary: More rst syntax fixes

Reviewed By: davidsonic

Differential Revision: D40977328

fbshipit-source-id: a3a3accbf2ba7cd9c84a0a82d0265010764a9d61
This commit is contained in:
Jeremy Reizenstein 2022-11-29 03:13:40 -08:00 committed by Facebook GitHub Bot
parent 94f321fa3d
commit a2c6af9250
14 changed files with 167 additions and 174 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
pytorch3d/io/pluggable.py
View File

@ -25,19 +25,16 @@ This module has the master functions for loading and saving data.
The main usage is via the IO object, and its methods
`load_mesh`, `save_mesh`, `load_pointcloud` and `save_pointcloud`.
For example, to load a mesh you might do
```
from pytorch3d.io import IO
For example, to load a mesh you might do::
mesh = IO().load_mesh("mymesh.obj")
```
from pytorch3d.io import IO
and to save a point cloud you might do
mesh = IO().load_mesh("mymesh.obj")
```
pcl = Pointclouds(...)
IO().save_pointcloud(pcl, "output_pointcloud.obj")
```
and to save a point cloud you might do::
pcl = Pointclouds(...)
IO().save_pointcloud(pcl, "output_pointcloud.obj")
"""

52
pytorch3d/io/ply_io.py
View File

@ -1067,33 +1067,33 @@ def load_ply(
is to use the IO.load_mesh and IO.load_pointcloud functions,
which can read more of the data.
Example .ply file format:
Example .ply file format::
ply
format ascii 1.0 { ascii/binary, format version number }
comment made by Greg Turk { comments keyword specified, like all lines }
comment this file is a cube
element vertex 8 { define "vertex" element, 8 of them in file }
property float x { vertex contains float "x" coordinate }
property float y { y coordinate is also a vertex property }
property float z { z coordinate, too }
element face 6 { there are 6 "face" elements in the file }
property list uchar int vertex_index { "vertex_indices" is a list of ints }
end_header { delimits the end of the header }
0 0 0 { start of vertex list }
0 0 1
0 1 1
0 1 0
1 0 0
1 0 1
1 1 1
1 1 0
4 0 1 2 3 { start of face list }
4 7 6 5 4
4 0 4 5 1
4 1 5 6 2
4 2 6 7 3
4 3 7 4 0
ply
format ascii 1.0 { ascii/binary, format version number }
comment made by Greg Turk { comments keyword specified, like all lines }
comment this file is a cube
element vertex 8 { define "vertex" element, 8 of them in file }
property float x { vertex contains float "x" coordinate }
property float y { y coordinate is also a vertex property }
property float z { z coordinate, too }
element face 6 { there are 6 "face" elements in the file }
property list uchar int vertex_index { "vertex_indices" is a list of ints }
end_header { delimits the end of the header }
0 0 0 { start of vertex list }
0 0 1
0 1 1
0 1 0
1 0 0
1 0 1
1 1 1
1 1 0
4 0 1 2 3 { start of face list }
4 7 6 5 4
4 0 4 5 1
4 1 5 6 2
4 2 6 7 3
4 3 7 4 0
Args:
f: A binary or text file-like object (with methods read, readline,

5
pytorch3d/ops/points_to_volumes.py
View File

@ -208,8 +208,8 @@ def add_pointclouds_to_volumes(
of `initial_volumes` with its `features` and `densities` updated with the
result of the pointcloud addition.
Example:
```
Example::
# init a random point cloud
pointclouds = Pointclouds(
points=torch.randn(4, 100, 3), features=torch.rand(4, 100, 5)
@ -229,7 +229,6 @@ def add_pointclouds_to_volumes(
initial_volumes=initial_volumes,
mode="trilinear",
)
```
Args:
pointclouds: Batch of 3D pointclouds represented with a `Pointclouds`

6
pytorch3d/renderer/implicit/harmonic_embedding.py
View File

@ -21,8 +21,8 @@ class HarmonicEmbedding(torch.nn.Module):
(i.e. vector along the last dimension) in `x`
into a series of harmonic features `embedding`,
where for each i in range(dim) the following are present
in embedding[...]:
```
in embedding[...]::
[
sin(f_1*x[..., i]),
sin(f_2*x[..., i]),
@ -34,7 +34,7 @@ class HarmonicEmbedding(torch.nn.Module):
cos(f_N * x[..., i]),
x[..., i], # only present if append_input is True.
]
```
where N corresponds to `n_harmonic_functions-1`, and f_i is a scalar
denoting the i-th frequency of the harmonic embedding.

12
pytorch3d/renderer/implicit/raymarching.py
View File

@ -25,20 +25,20 @@ class EmissionAbsorptionRaymarcher(torch.nn.Module):
(i.e. its density -> 1.0).
EA first utilizes `rays_densities` to compute the absorption function
along each ray as follows:
```
along each ray as follows::
absorption = cumprod(1 - rays_densities, dim=-1)
```
The value of absorption at position `absorption[..., k]` specifies
how much light has reached `k`-th point along a ray since starting
its trajectory at `k=0`-th point.
Each ray is then rendered into a tensor `features` of shape `(..., feature_dim)`
by taking a weighed combination of per-ray features `rays_features` as follows:
```
by taking a weighed combination of per-ray features `rays_features` as follows::
weights = absorption * rays_densities
features = (rays_features * weights).sum(dim=-2)
```
Where `weights` denote a function that has a strong peak around the location
of the first surface point that a given ray passes through.

33
pytorch3d/renderer/implicit/raysampling.py
View File

@ -32,23 +32,22 @@ class MultinomialRaysampler(torch.nn.Module):
have uniformly-spaced z-coordinates between a predefined
minimum and maximum depth.
The raysampler first generates a 3D coordinate grid of the following form:
```
/ min_x, min_y, max_depth -------------- / max_x, min_y, max_depth
/ /|
/ / | ^
/ min_depth min_depth / | |
min_x ----------------------------- max_x | | image
min_y min_y | | height
| | | |
| | | v
| | |
| | / max_x, max_y, ^
| | / max_depth /
min_x max_y / / n_pts_per_ray
max_y ----------------------------- max_x/ min_depth v
< --- image_width --- >
```
The raysampler first generates a 3D coordinate grid of the following form::
/ min_x, min_y, max_depth -------------- / max_x, min_y, max_depth
/ /|
/ / | ^
/ min_depth min_depth / | |
min_x ----------------------------- max_x | | image
min_y min_y | | height
| | | |
| | | v
| | |
| | / max_x, max_y, ^
| | / max_depth /
min_x max_y / / n_pts_per_ray
max_y ----------------------------- max_x/ min_depth v
< --- image_width --- >
In order to generate ray points, `MultinomialRaysampler` takes each 3D point of
the grid (with coordinates `[x, y, depth]`) and unprojects it

58
pytorch3d/renderer/implicit/renderer.py
View File

@ -41,13 +41,13 @@ class ImplicitRenderer(torch.nn.Module):
as well as the `volumetric_function` `Callable`, which defines a field of opacity
and feature vectors over the 3D domain of the scene.
A standard `volumetric_function` has the following signature:
```
def volumetric_function(
ray_bundle: Union[RayBundle, HeterogeneousRayBundle],
**kwargs,
) -> Tuple[torch.Tensor, torch.Tensor]
```
A standard `volumetric_function` has the following signature::
def volumetric_function(
ray_bundle: Union[RayBundle, HeterogeneousRayBundle],
**kwargs,
) -> Tuple[torch.Tensor, torch.Tensor]
With the following arguments:
`ray_bundle`: A RayBundle or HeterogeneousRayBundle object
containing the following variables:
@ -79,32 +79,32 @@ class ImplicitRenderer(torch.nn.Module):
Example:
A simple volumetric function of a 0-centered
RGB sphere with a unit diameter is defined as follows:
```
def volumetric_function(
ray_bundle: Union[RayBundle, HeterogeneousRayBundle],
**kwargs,
) -> Tuple[torch.Tensor, torch.Tensor]:
RGB sphere with a unit diameter is defined as follows::
# first convert the ray origins, directions and lengths
# to 3D ray point locations in world coords
rays_points_world = ray_bundle_to_ray_points(ray_bundle)
def volumetric_function(
ray_bundle: Union[RayBundle, HeterogeneousRayBundle],
**kwargs,
) -> Tuple[torch.Tensor, torch.Tensor]:
# set the densities as an inverse sigmoid of the
# ray point distance from the sphere centroid
rays_densities = torch.sigmoid(
-100.0 * rays_points_world.norm(dim=-1, keepdim=True)
)
# first convert the ray origins, directions and lengths
# to 3D ray point locations in world coords
rays_points_world = ray_bundle_to_ray_points(ray_bundle)
# set the ray features to RGB colors proportional
# to the 3D location of the projection of ray points
# on the sphere surface
rays_features = torch.nn.functional.normalize(
rays_points_world, dim=-1
) * 0.5 + 0.5
# set the densities as an inverse sigmoid of the
# ray point distance from the sphere centroid
rays_densities = torch.sigmoid(
-100.0 * rays_points_world.norm(dim=-1, keepdim=True)
)
# set the ray features to RGB colors proportional
# to the 3D location of the projection of ray points
# on the sphere surface
rays_features = torch.nn.functional.normalize(
rays_points_world, dim=-1
) * 0.5 + 0.5
return rays_densities, rays_features
return rays_densities, rays_features
```
"""
def __init__(self, raysampler: Callable, raymarcher: Callable) -> None:

12
pytorch3d/renderer/implicit/utils.py
View File

@ -73,13 +73,13 @@ def ray_bundle_to_ray_points(
extending each ray according to the corresponding length.
E.g. for 2 dimensional tensors `ray_bundle.origins`, `ray_bundle.directions`
and `ray_bundle.lengths`, the ray point at position `[i, j]` is:
```
and `ray_bundle.lengths`, the ray point at position `[i, j]` is::
ray_bundle.points[i, j, :] = (
ray_bundle.origins[i, :]
+ ray_bundle.directions[i, :] * ray_bundle.lengths[i, j]
)
```
Note that both the directions and magnitudes of the vectors in
`ray_bundle.directions` matter.
@ -109,13 +109,13 @@ def ray_bundle_variables_to_ray_points(
ray length:
E.g. for 2 dimensional input tensors `rays_origins`, `rays_directions`
and `rays_lengths`, the ray point at position `[i, j]` is:
```
and `rays_lengths`, the ray point at position `[i, j]` is::
rays_points[i, j, :] = (
rays_origins[i, :]
+ rays_directions[i, :] * rays_lengths[i, j]
)
```
Note that both the directions and magnitudes of the vectors in
`rays_directions` matter.

32
pytorch3d/renderer/opengl/opengl_utils.py
View File

@ -285,26 +285,26 @@ class _DeviceContextStore:
The EGL/CUDA contexts are not meant to be created and destroyed all the time,
and having multiple on a single device can be troublesome. Intended use is entirely
transparent to the user:
transparent to the user::
```
rasterizer1 = MeshRasterizerOpenGL(...some args...)
mesh1 = load_mesh_on_cuda_0()
rasterizer1 = MeshRasterizerOpenGL(...some args...)
mesh1 = load_mesh_on_cuda_0()
# Now rasterizer1 will request EGL/CUDA contexts from global_device_context_store
# on cuda:0, and since there aren't any, the store will create new ones.
rasterizer1.rasterize(mesh1)
# Now rasterizer1 will request EGL/CUDA contexts from
# global_device_context_store on cuda:0, and since there aren't any, the
# store will create new ones.
rasterizer1.rasterize(mesh1)
# rasterizer2 also needs EGL & CUDA contexts. But global_context_store already has
# them for cuda:0. Instead of creating new contexts, the store will tell rasterizer2
# to use them.
rasterizer2 = MeshRasterizerOpenGL(dcs)
rasterize2.rasterize(mesh1)
# rasterizer2 also needs EGL & CUDA contexts. But global_context_store
# already has them for cuda:0. Instead of creating new contexts, the store
# will tell rasterizer2 to use them.
rasterizer2 = MeshRasterizerOpenGL(dcs)
rasterize2.rasterize(mesh1)
# When rasterizer1 needs to render on cuda:1, the store will create new contexts.
mesh2 = load_mesh_on_cuda_1()
rasterizer1.rasterize(mesh2)
# When rasterizer1 needs to render on cuda:1, the store will create new contexts.
mesh2 = load_mesh_on_cuda_1()
rasterizer1.rasterize(mesh2)
```
"""
def __init__(self):

1
pytorch3d/renderer/points/pulsar/renderer.py
View File

@ -11,7 +11,6 @@ Proper Python support for pytorch requires creating a torch.autograd.function
here and a torch.nn.Module is exposed for the use in more complex models.
"""
import logging
import math
import warnings
from typing import Optional, Tuple, Union

78
pytorch3d/structures/volumes.py
View File

@ -97,12 +97,12 @@ class Volumes:
in the world coordinates.
- They are specified with the following mapping that converts
points `x_local` in the local coordinates to points `x_world`
in the world coordinates:
```
x_world = (
x_local * (volume_size - 1) * 0.5 * voxel_size
) - volume_translation,
```
in the world coordinates::
x_world = (
x_local * (volume_size - 1) * 0.5 * voxel_size
) - volume_translation,
here `voxel_size` specifies the size of each voxel of the volume,
and `volume_translation` is the 3D offset of the central voxel of
the volume w.r.t. the origin of the world coordinate frame.
@ -110,12 +110,12 @@ class Volumes:
the world coordinate units. `volume_size` is the spatial size of
the volume in form of a 3D vector `[width, height, depth]`.
- Given the above definition of `x_world`, one can derive the
inverse mapping from `x_world` to `x_local` as follows:
```
x_local = (
(x_world + volume_translation) / (0.5 * voxel_size)
) / (volume_size - 1)
```
inverse mapping from `x_world` to `x_local` as follows::
x_local = (
(x_world + volume_translation) / (0.5 * voxel_size)
) / (volume_size - 1)
- For a trivial volume with `volume_translation==[0, 0, 0]`
with `voxel_size=-1`, `x_world` would range
from -(volume_size-1)/2` to `+(volume_size-1)/2`.
@ -139,13 +139,13 @@ class Volumes:
to `x_local=(-1, 0, 1)`.
- For a "trivial" volume `v` with `voxel_size = 1.`,
`volume_translation=[0., 0., 0.]`, the following holds:
```
torch.nn.functional.grid_sample(
v.densities(),
v.get_coord_grid(world_coordinates=False),
align_corners=True,
) == v.densities(),
```
torch.nn.functional.grid_sample(
v.densities(),
v.get_coord_grid(world_coordinates=False),
align_corners=True,
) == v.densities(),
i.e. sampling the volume at trivial local coordinates
(no scaling with `voxel_size`` or shift with `volume_translation`)
results in the same volume.
@ -588,12 +588,12 @@ class VolumeLocator:
in the world coordinates.
- They are specified with the following mapping that converts
points `x_local` in the local coordinates to points `x_world`
in the world coordinates:
```
x_world = (
x_local * (volume_size - 1) * 0.5 * voxel_size
) - volume_translation,
```
in the world coordinates::
x_world = (
x_local * (volume_size - 1) * 0.5 * voxel_size
) - volume_translation,
here `voxel_size` specifies the size of each voxel of the volume,
and `volume_translation` is the 3D offset of the central voxel of
the volume w.r.t. the origin of the world coordinate frame.
@ -601,12 +601,12 @@ class VolumeLocator:
the world coordinate units. `volume_size` is the spatial size of
the volume in form of a 3D vector `[width, height, depth]`.
- Given the above definition of `x_world`, one can derive the
inverse mapping from `x_world` to `x_local` as follows:
```
x_local = (
(x_world + volume_translation) / (0.5 * voxel_size)
) / (volume_size - 1)
```
inverse mapping from `x_world` to `x_local` as follows::
x_local = (
(x_world + volume_translation) / (0.5 * voxel_size)
) / (volume_size - 1)
- For a trivial volume with `volume_translation==[0, 0, 0]`
with `voxel_size=-1`, `x_world` would range
from -(volume_size-1)/2` to `+(volume_size-1)/2`.
@ -629,14 +629,14 @@ class VolumeLocator:
`DxHxW = 5x5x5`, the point `x_world = (-2, 0, 2)` gets mapped
to `x_local=(-1, 0, 1)`.
- For a "trivial" volume `v` with `voxel_size = 1.`,
`volume_translation=[0., 0., 0.]`, the following holds:
```
torch.nn.functional.grid_sample(
v.densities(),
v.get_coord_grid(world_coordinates=False),
align_corners=True,
) == v.densities(),
```
`volume_translation=[0., 0., 0.]`, the following holds::
torch.nn.functional.grid_sample(
v.densities(),
v.get_coord_grid(world_coordinates=False),
align_corners=True,
) == v.densities(),
i.e. sampling the volume at trivial local coordinates
(no scaling with `voxel_size`` or shift with `volume_translation`)
results in the same volume.

23
pytorch3d/transforms/math.py
View File

@ -22,18 +22,17 @@ def acos_linear_extrapolation(
domain of `(-1, 1)`. This allows for stable backpropagation in case `x`
is not guaranteed to be strictly within `(-1, 1)`.
More specifically:
```
bounds=(lower_bound, upper_bound)
if lower_bound <= x <= upper_bound:
acos_linear_extrapolation(x) = acos(x)
elif x <= lower_bound: # 1st order Taylor approximation
acos_linear_extrapolation(x)
= acos(lower_bound) + dacos/dx(lower_bound) * (x - lower_bound)
else: # x >= upper_bound
acos_linear_extrapolation(x)
= acos(upper_bound) + dacos/dx(upper_bound) * (x - upper_bound)
```
More specifically::
bounds=(lower_bound, upper_bound)
if lower_bound <= x <= upper_bound:
acos_linear_extrapolation(x) = acos(x)
elif x <= lower_bound: # 1st order Taylor approximation
acos_linear_extrapolation(x)
= acos(lower_bound) + dacos/dx(lower_bound) * (x - lower_bound)
else: # x >= upper_bound
acos_linear_extrapolation(x)
= acos(upper_bound) + dacos/dx(upper_bound) * (x - upper_bound)
Args:
x: Input `Tensor`.

6
pytorch3d/transforms/transform3d.py
View File

@ -256,12 +256,12 @@ class Transform3d:
The conversion from the 4x4 SE(3) matrix `transform` to the
6D representation `log_transform = [log_translation | log_rotation]`
is done as follows:
```
is done as follows::
log_transform = log(transform.get_matrix())
log_translation = log_transform[3, :3]
log_rotation = inv_hat(log_transform[:3, :3])
```
where `log` is the matrix logarithm
and `inv_hat` is the inverse of the Hat operator [2].

6
pytorch3d/utils/camera_conversions.py
View File

@ -35,10 +35,10 @@ def cameras_from_opencv_projection(
to the NDC screen convention of PyTorch3D.
More specifically, the OpenCV convention projects points to the OpenCV screen
space as follows:
```
space as follows::
x_screen_opencv = camera_matrix @ (R @ x_world + tvec)
```
followed by the homogenization of `x_screen_opencv`.
Note: